Sign up to see the future, today<\/p>\n
Can\u2019t-miss innovations from the bleeding edge of science and tech<\/p>\n
David Foster Wallace once asked readers to \u201cconsider the lobster,\u201d in his famous essay about the ethics of boiling the creatures alive.<\/p>\n
But new tech brings new horrors, and perhaps Wallace, if he were alive today, would instead be writing about biohacking lobster brains so they can be taken over by an AI agent. Because that\u2019s exactly what a group of young tech founders in San Francisco have attempted to do, according to a riveting side plot in a new feature from The Atlantic<\/a> about \u201chacker houses\u201d in the Bay Area. And what happened tells you everything you need to know about the ethics \u2014 and actual ability, perhaps \u2014 of the people trying to climb their way to the top of this multibillion dollar industry.<\/p>\n Our two would-be Doctor Lobstersteins are 32-year-old Elliot Roth and 19-year-old William Joy. The duo live in one of the many so-called hacker houses that are popping up across the city: college dorm-esque refuges where young tech entrepreneurs bunk together and bounce wild ideas off each other. Theirs is called Biopunk House. It\u2019s home to 20 aspiring tech overlords, and it\u2019s part of a broader network called the Residency, which OpenAI CEO Sam Altman<\/a> is an adviser on.<\/p>\n The duo\u2019s plan was to use a kit designed for remote-controling a cockroach and implanting it in the lobsters. From there, after demonstrating that they could use to it direct the lobster to do basic stuff like pinch its claws, they\u2019d hook up the now cyborg crustacean to OpenClaw<\/a>, the wildly popular open source AI agent that has a lobster as its logo. (Yes, this toying with a living being\u2019s biology is also a little joke.)<\/p>\n But wait, aren\u2019t lobster brains and roach brains way different from each other? And wouldn\u2019t this require some serious neurosurgical skills? Don\u2019t worry about it. If the rise of \u201cvibe-coding<\/a>\u201d has taught us anything about tech bros, it\u2019s that they don\u2019t treat the question of personal ability seriously. If the belief in the technology is there, it\u2019ll all somehow work out.<\/p>\n \u201cI\u2019m pretty sure it\u2019s going to be the first real instance of a complex AI agent interfacing with a biological organism,\u201d Joy rhapsodized to The Atlantic reporter during the planning stages of the experiment.<\/p>\n Don\u2019t get the wrong idea, the pair insisted. Roth and Joy said they were deeply concerned about the lobsters\u2019 welfare, which is why they considered administering them with some kind of anesthetic.<\/p>\n \u201cWe are going to give a lot of thought to, \u2018How can we ensure that they don\u2019t suffer?’\u201d Joy told the magazine, noting that the animals were \u201calready destined for the dinner table.\u201d (They would eat the lobsters afterwards, they promised, to honor their contributions to bro science.)<\/p>\n Shockingly, their ambitions hit a brutal reality check \u2014 for the lobsters mostly, since the humans involved here made out pretty okay. When the reporter checked in on the lobster experiment weeks later, he saw that the tank where the lobsters had been housed \u201cwas conspicuously empty of any life or, for that matter, liquid.\u201d<\/p>\n That\u2019s because the lobsters died before the surgery could even happen. The creatures did go under the knife \u2014 as part of Joy\u2019s \u201cpreliminary\u201d autopsy. The young man said he\u2019d been going through an ethical crisis wondering \u201cwhether I should even be doing this.\u201d<\/p>\n And no, the lobsters were not eaten as promised.<\/p>\n What did Roth and Joy hope to gain from biohacking the lobsters? Did they stop for a second and think about if they had anywhere near the level of expertise to pull this off? What were they hoping to prove, other than that it was merely possible? We can\u2019t know what\u2019s going on in their minds anymore than our surgeons manqu\u00e9 can know the nuances of a lobster\u2019s.\u00a0<\/p>\n But this epitomizes the flippancy that those in AI have towards ethics, and maybe even towards the real world beyond their hermetic hype-doms of tech evangelism. They were toying with these animals\u2019 wellbeing. And when it came time to actually put their nose to the grindstone and attempt a serious feat of neurosurgery, they backed out. The lobsters died anyway out of negligence, with Joy admitting that they might\u2019ve gotten the salinity of the water wrong.\u00a0These were serious moral questions and basic operational details to iron out, but Roth and Joy seemed to sincerely believed they could vibe their way through it.<\/p>\n This is also the attitude of the AI industry at large. Why are AI chatbots still leading people down destructive delusions and mental health spirals? What will AI companies do to meaningfully clamp down on all the harmful material they\u2019re generating, from misinformation to nonconsensual deepfakes and instructions for hacking? And what about its environmental impact and its monstrous energy demands? The answer to all of these is to press on, to prompt through it, and occasionally feel guilty or worried about it in the way that Anthropic \u2014 to give one prime example \u2014 constantly acts like it does<\/a>.<\/p>\n